AI · GroupNucleus Technology’s Three Attempts to List in Hong Kong: Is It Just to Solve Urgent Funding Needs?

Produced by | Frontline of Entrepreneurship

Author | Yu Ying

Editor | Egg Boss

Design | Xing Jing

Review | Song Wen

After “Hangzhou’s Six Little Dragons” such as DeepSeek and Yushu Technology successively sparked discussions with breakthrough technologies, the “Six Little Dragons” have gradually become a clear label, carrying external expectations.

Among them, GroupNucleus Technology may be the most low-key and unique. It doesn’t have the halo of large models or the coolness of robots, but it holds a product almost all home interior designers know—Kujiale.

On February 24, GroupNucleus Technology filed for listing in Hong Kong for the third time. From its failed U.S. IPO in 2021, to two unsuccessful Hong Kong IPO applications by 2025, and now attempting again, the road for GroupNucleus Technology’s IPO seems bumpy.

If this time succeeds, it will be the first among the “Six Little Dragons” to enter the capital market, and will earn the title of “Global Space Intelligence’s First Stock.”

But with the halo comes difficulty. The prospectus shows that this “space intelligence” company currently derives nearly 97% of its revenue from membership fees for home design software, with core customers being renovation companies, building material suppliers, and furniture manufacturers.

As the real estate industry enters a deep adjustment period, these “big spenders” are tightening budgets. Meanwhile, the company’s repeatedly emphasized new AI businesses—space intelligence and embodied intelligence training platforms—contribute only 3% of revenue.

On one side are the “Six Little Dragons” with their shining halo; on the other, growth anxiety under the real estate cycle. One side is high-margin tech business; the other is the reality of losses. What kind of company is GroupNucleus? How much room for development does it really have?

1. Revenue Growth Slowing, Home Decor Companies Still Core Customers

“Frontline of Entrepreneurship” notes that if you search for “useful home design apps” on social platforms or search engines, Kujiale is likely to be recommended, which is also GroupNucleus’s core product.

Launched in 2013, this cloud-based design software emphasizes quick drag-and-drop 3D design, real-time rendering, and BIM capabilities. Simply put, on Kujiale, designers upload CAD drawings or floor plans, and the system automatically recognizes structural information and generates editable 3D design schemes.

Compared to traditional cloud design software, Kujiale’s most prominent advantage is “speed.” It can render a design into a 5K high-definition image in just 10 seconds, achieving near real-time visualization.

Additionally, designers can send complete 3D data—including dimensions, materials, and specifications—directly to manufacturers. The system can automatically generate production drawings, reducing errors.

(Images / Kujiale platform works)

When Kujiale launched in 2013, it coincided with a peak in China’s real estate market. During this period, the company benefited from the real estate boom and developed rapidly.

According to financial data submitted during GroupNucleus’s U.S. listing, revenue in 2019 and 2020 was 282 million yuan and 353 million yuan, respectively, with a year-over-year growth of 25.2%. In the first quarter of 2021, revenue reached 101 million yuan, up 33.32%.

However, this growth trend has reversed in recent years. As the real estate industry entered a deep adjustment, from 2023 to 2025, the company’s revenue growth rate slowed from 13.7% to 8.6%. The latest prospectus shows revenues of approximately 664 million yuan, 755 million yuan, and 820 million yuan over these three years, with a clear slowdown.

The latest prospectus also explicitly states the core risks: the company’s customers are highly concentrated in the home furnishing and real estate upstream and downstream industries. A downturn in these sectors could directly impact customers’ willingness and ability to pay.

Facing this external environment, GroupNucleus is seeking new ways out, most notably by emphasizing its technological attributes.

For example, in 2024, leveraging Kujiale’s accumulation of massive 3D worlds, physically accurate, and computable spatial data, the company launched the SpatialVerse space intelligence solution.

It is understood that this is a training platform specifically for embodied intelligence. In reality, collecting real data for robots is costly—costs often start at hundreds of thousands of yuan—and efficiency is limited. The virtual environment built by SpatialVerse allows robots to undergo large-scale training iterations in digital worlds. Currently, clients include Zhiyuan Robot and Galaxy General.

(Images / GroupNucleus official website)

However, despite interesting technological layouts, the company’s revenue structure remains highly single. Latest data shows nearly 97% of revenue comes from software subscriptions (membership fees).

Breaking down these membership fees, 81.6% are from enterprise clients, mainly companies in the entire home furnishing industry chain, such as full-house customization brands, furniture factories, renovation companies, building material distributors, as well as chain stores and commercial space renovation firms.

GroupNucleus sells annual subscription accounts to these enterprises, with prices ranging from a few thousand yuan for basic versions to hundreds of thousands or even over a million yuan for customized packages.

By 2025, the company aims to have over 47,000 paying enterprise customers, with 424 top clients paying over 200,000 yuan annually, each contributing an average of 856,000 yuan per year—forming the most stable revenue pillar.

(Images / GroupNucleus IPO prospectus)

(Images / GroupNucleus IPO prospectus)

Therefore, it can be said that the core financial supporters of GroupNucleus are still enterprise clients in the home furnishing and renovation industry.

Additionally, 15.3% of the company’s membership fees come from individual users—ordinary homeowners, independent designers, and design enthusiasts who are renovating their homes. The remaining 3.1% are technical service fees, amounting to only 25.2 million yuan in 2025.

Thus, despite having laid out in space intelligence and AI training platforms, these businesses are still in early stages, contributing little to overall revenue and not yet forming a second growth curve.

(Images / GroupNucleus IPO prospectus)

It is clear that for a company attempting to drive growth through technological innovation, the key challenge remains: how to shift from “storytelling” to a true structural transformation.

2. Over 80% Gross Margin, Yet Difficult to Make Money

From a revenue model perspective, in the coming years, GroupNucleus is likely to continue relying mainly on membership fees from home furnishing companies. Under this business logic, the company’s advantages and disadvantages are quite evident.

First, due to asset-light operations, GroupNucleus’s gross margin exceeds 80%, a relatively high level in the industry. This is one reason many investors are attracted—high gross margins suggest ample profit space in core business, with potential for profitability.

(Images / GroupNucleus IPO prospectus)

However, high gross margin has not translated into overall profitability, which is the most prominent shortcoming of the business model.

From 2023 to 2025, the company is expected to incur losses of approximately 646 million yuan, 513 million yuan, and 428 million yuan, totaling a cumulative loss of 1.587 billion yuan over three years. Adjusted net profit is only expected to turn positive in 2025, reaching 57.1 million yuan.

So, if the core business has an 80% gross margin and is profitable, why has GroupNucleus been in long-term loss? The answer is similar to many online platforms: it’s a “burning money for growth” model, with large and ongoing expenses.

Specifically, the main reasons for losses are high R&D and marketing costs.

As a technology-driven company, GroupNucleus understands the importance of innovation and product development, investing heavily in R&D to maintain technological leadership and continuous innovation.

The prospectus states that losses mainly stem from sustained large investments in product development, technical support, and marketing. From 2023 to 2025, R&D expenditure alone exceeds 1 billion yuan.

(Images / GroupNucleus IPO prospectus)

Meanwhile, to expand market share and enhance brand influence, the company also invests heavily in marketing. From 2023 to 2025, sales and marketing expenses account for 53.7%, 43.2%, and 33.4% of revenue, respectively—huge costs for any enterprise.

What is the effect of this spending? User data shows that the “burning money for growth” strategy has indeed increased scale, but growth quality is questionable.

By 2025, the company’s software attracts an average of about 2.5 million monthly active users. Personal users increased from 390,000 in 2023 to 426,800 in 2024, then slightly declined to 416,100 in 2025. Enterprise users grew from 41,000 in 2023 to 47,000 in 2025.

However, customer retention and willingness to pay continue to decline. Net revenue retention for enterprise clients fell from 106% in 2023 to 100.7%; for individual users, it plummeted from 106.5% to 86.4%.

Meanwhile, deferred revenue (prepaid membership fees), which indicates future payment confidence, decreased from 617 million yuan in 2023 to 510 million yuan in 2025. Customers are shifting from long-term subscriptions to annual contracts, reflecting a generally cautious industry outlook.

(Images / GroupNucleus IPO prospectus)

While user numbers are increasing, retention rates are dropping, indicating that revenue growth depends on “new customer acquisition” rather than “retaining old customers.” This forces the company to continually spend to attract new clients, with marginal customer acquisition costs rising, creating a cycle of “more growth, more spending.”

The weakening willingness to pay also shows that the downturn in real estate first impacts small renovation companies, building material distributors, and mom-and-pop shops—clients with weaker risk resistance. As their business shrinks, they cut software budgets, downgrade packages, reduce accounts, or stop renewing.

Thus, under this business model, the company is caught in a “high gross margin but loss-making” trap.

The most direct way to break this cycle is to cut costs. By 2025, GroupNucleus reduced sales and marketing expenses by 16% to 274 million yuan, and R&D expenses also decreased by 13.6% to 291 million yuan, related to its technological moat.

With cost cuts, the company finally turned profitable, but whether this “cost-cutting” approach is sustainable remains to be seen.

It is worth noting that one of the goals of this IPO is international expansion. Since 2018, GroupNucleus launched the international version Coohom, mainly targeting South Korea, Southeast Asia, India, the U.S., and Japan. During the reporting period, international revenue accounted for 6.0%, 7.6%, and 9.0% of total revenue.

But in this money-burning mode, will overseas expansion create new hope or repeat the domestic pattern of “high investment, low return”? The answer remains to be seen, depending on time and market response.

3. High Asset-Liability Ratio, Multiple Listing Attempts for “Capital Injection”

Besides profitability issues, large ongoing expenses have kept GroupNucleus’s cash flow under pressure, and operational “bleeding” has not been fundamentally addressed.

From 2023 to 2025, net cash flow from operating activities was -62.57 million yuan, -108 million yuan, and -19.22 million yuan, respectively—three consecutive years of negative cash flow.

This means the core business cannot generate positive cash flow and must continuously consume existing funds, indicating weak “self-sustaining” ability and reliance on external capital to fill the gap.

(Images / GroupNucleus IPO prospectus)

Moreover, the company’s debt level is high. As of the end of 2025, total liabilities reached 4.735 billion yuan, while total assets were only 551 million yuan; current liabilities totaled 4.672 billion yuan, with current assets only 420 million yuan.

Even after excluding the 4.091 billion yuan of redemption liabilities, the asset-liability ratio remains high at 116.87%, facing insolvency risk.

(Images / GroupNucleus IPO prospectus)

More concerning is the decreasing reserve funds for immediate debt repayment. At each reporting period, cash and cash equivalents were 366 million yuan, 448 million yuan, and 357 million yuan, respectively, with a noticeable drop in 2025 compared to 2024.

“Frontline of Entrepreneurship” notes that under operational pressure, GroupNucleus has conducted multiple financing rounds. Data shows that from 2013 onward, it has completed 8 rounds of financing totaling over $430 million.

Its shareholder roster is quite star-studded, including Xiaomi’s Lei Jun-controlled Shunwei Capital, Hillhouse Capital, IDG Capital, GGV Jiyuan Capital, and Matrix Partners.

Media reports indicate that in 2021, during its push into the U.S. stock market, the company’s valuation reached $2 billion (about 14.5 billion RMB).

However, investor enthusiasm has waned. The last external funding was in 2021, and no further financing has been secured since.

This suggests the company’s appeal to capital has diminished. Without timely “capital injection” from other channels, and given current costs, maintaining normal operations will be difficult, and the risk of a capital chain break increases.

This is the third time GroupNucleus has filed for listing in Hong Kong, having previously abandoned plans for a U.S. IPO. This raises questions about its core motivation: is it to raise funds to alleviate urgent debt and cash flow issues, or simply a routine financing for healthy business development?

More critically, if the IPO does not raise enough funds, what specific measures will the company take to address short-term debt risks and ensure ongoing operations, protecting the interests of minority shareholders? “Frontline of Entrepreneurship” attempted to seek further clarification from GroupNucleus but has not received a response before publication.

Overall, despite the dual halos of “Hangzhou’s Six Little Dragons” and “Global Space Intelligence’s First Stock,” the company’s three IPO attempts cannot escape the growth constraints of its home industry, and new business implementation is still distant.

Next, GroupNucleus’s turnaround may depend not only on opening the capital market door but also on truly breaking free from industry dependence, transforming technological layouts into sustainable growth drivers.

Note: The lead image in the article is from Jiemian Gallery.